Search

The development of aberrant cell proliferation in the lungs is a problematic condition that has the potential to result in death. On the list of diseases that most frequently result in mortality, lung cancer takes first place. The early stages of lung cancer are notoriously difficult to diagnose due to the fact that cancer cells with dimensions less than very small are notoriously difficult to spot by imaging. If the cell abnormalities are discovered in the early stages, it will be possible to begin therapy sooner, which will result in an improved chance of the patient surviving the illness. Several different image processing strategies can be utilized in the diagnostic phase of patient care to help spot signs of disease. In this paper, classification of Lung Cancer from chest X-ray images has been done using optimized Convolutional Neural Network (OCNN) and Ant Lion Optimization (ALO) algorithm. In pre-processing step, the contrast of all images are enhanced using Histogram Equalization (HE) method and the noises are removed from all images using Median Filtering. After the pre-processing step, feature extraction is performed using Gray Level Spatial Dependence (GLSD) to extract the statistical features. The feature vector is then trained and classified using OCNN-ALO algorithm. The ALO algorithm is used to optimize the hyper parameters of CNN layers. It classifies the lung images into normal and lung tumor affected. Performance results have indicated that OCNN-ALO attains the superior performance with 95.15% accuracy, 85.43% sensitivity, 93.4% specificity and 76.43% F1-score. [ABSTRACT FROM AUTHOR]

This paper discusses the application of the Content-Based Image Retrieval (CBIR) system in image classification for various types of batik in the region. The aim is to detect and recognize certain types of batik with high accuracy, as well as develop a digital image-based security system for various types of batik with various motifs, patterns, and colors. In this research, a typical South Sulawesi batik type, namely Lontara motif batik, was used as a sample. The methods used to perform feature extraction and classification are the Gray Level Co-Occurrence Matrix (GLCM) and Multilayer Perceptron (MLP). The use of GLCM is equipped with the Angular Second Moment (ASM) feature with contrast, correlation, energy, and homogeneity parameters, so that the image data extraction results will be more accurate. Apart from that, the combination with the Multi-Layer Perceptron (MLP) algorithm will make it easier for the system to carry out classification because MLP uses three training functions, namely Levenberg-Marquardt (LM), Variable Learning Rate Backpropagation (GDX), and Scaled Conjugate Gradient (SCG), which produce high accuracy values at smaller epochs. The best accuracy value obtained in the test results was 88.89%. The combination of the GLCM-MLP method can be used as a solution to accurately detect various batik motifs. [ABSTRACT FROM AUTHOR]

In Handwritten Character Recognition (HCR), interest in feature extraction has been on the increase with the abundance of algorithms derived to increase the accuracy of classification. In this paper, a metaheuristic approach for feature extraction technique in HCR based on Crow Search Algorithm (CSA) was proposed. Freeman Chain Code (FCC) was used as data representation. The main problem in representing a character using FCC is that the results of the extractions depend on the starting points that affected the route length of chain code. To solve this problem, the metaheuristic approach via CSA was proposed to find the shortest route length and minimum computational time for HCR. The performance measurements of the proposed CS-FCC extraction algorithm are the route lengths and computation times. The experiments on the algorithms are performed based on the chain code representation derived from established previous works of Center of Excellence for Document Analysis and Recognition (CEDAR) dataset which consists of 126 upper-case letter characters. Based on the result, the proposed CS-FCC obtained 1880.28 in term of route length and only needs 1.10 second to solve the whole character images. [ABSTRACT FROM AUTHOR]

To detect the objects around an autonomous vehicle is very essential to operate safely. This paper presents to detect and classify the objects for assisting autonomous driving. In autonomous driving systems, the task of object detection itself is one of the most important prerequisites to autonomous navigation. Deep learning one of the computer vision tasks, perform object detection very effectively than compared to earlier methods and this project is to detect the objects like vehicles, persons, traffic lights, etc. In this work, an approach to object detection in deep learning that makes the bounding box for an image to predict is explored. Object detection is the method of detecting the objects present in a given image. Apart from detecting the number of objects present in an image it also specifies in which location that object is present in the image. The objects are detected by means of bounding boxes. In the existing system algorithm like Convolutional Neural Network (CNN) using Resnet-50 were used to detect the objects like vehicles, persons, traffic lights separately. The problem identified here is, in the existing system the camera is fixed in a particular place and it detects objects only if the objects come into the camera frames. It is not detecting both objects and lanes simultaneously when the autonomous vehicle is in motion at any location. To overcome these problems, in the proposed system object detection is performed by mounting camera in front of the moving vehicle. You Only Look Once (YOLO) V3 Algorithm is used for the process of object detection. Compared to earlier detection approaches YOLO V3 shows improvement in detection accuracy. It provides good feature extraction and detection in large-scale. The proposed YOLO al gorithm has better average precision value for detecting all objects than compared to existing CNN using Resnet-50. In addition YOLO V3 algorithm performs lane detection apart from object detection. [ABSTRACT FROM AUTHOR]

Sentiment analysis of text data has become very much popular in past few years, because of the interesting challenges it can offer. Among these challenges, sarcasm is unique. We can address sarcasm as the 'Achilles heel ' of sentiment analysis. Detection of sarcasm from the given data is same time complicated and interesting. It is interesting because, if researchers can find the optimized solutions for finding sarcastic words in the data, it will enhance the sentiment analysis of that data. Even humans also have difficulties in understanding the actual interpretation of a sentence, if it is presented indirectly. This means, a person stated something, but he meant contradictory to the word meaning of the statement. This makes sarcasm detection an interesting topic for researchers. In this paper, we evaluated the performance of several machine learning models like Support vector machine, Naïve Bayes, Decision tree etc., and different ensemble models like Random Forest, XGBoost and AdaBoost etc., with the collaboration of various feature extraction methods such as Term Frequency-Inverse Document Frequency etc. The main evaluation metrics we used to evaluate the performance are accuracy, precision, f-score and recall. Based on the results, we concluded that the models such as XGBoost, LightGBM and Bagging classifiers provide better results in detecting sarcasm with respect to other machine learning models. [ABSTRACT FROM AUTHOR]

Bearing is a critical component in the condition monitoring of any rotary machines. In the rotating machines, the bearing is the most widely used rotating elements. Any fault on the surface of the ball bearing will cause surface fatigue and leads to the failure of the ball bearing. They are caused mainly by the cyclic loading of the shaft and it is an inevitable one, when the bearing is in running condition. By condition monitoring these bearings it will helps us to proliferate the performance of the machines and reduce the maintenance cost of the industries. This paper presents the uses of Fuzzy Unordered Rule Induction Algorithm (FURIA) classification for the features collected from the vibration signals of the fault bearings and the good ones. The features from the signal can be extracted by the J48 statistical feature extraction method. The analysis is achieved by comparing the feature extraction of signals obtained in both the conditions in LabVIEW. By applying the FURIA classification to the extracted features in the Waikato Environment for Knowledge Analysis (WEKA) classification software and the faults in the bearings are diagnosed. The results show that FURIA algorithm has a better classification accuracy of 95.83% for identifying the bearing conditions. [ABSTRACT FROM AUTHOR]

This paper provide a general overview on different kinds of mining categorical covid-19 data Using Chi-Square and Logistic Regression Algorithms. Our study methodology are categorized into three steps, the first one is data preprocessing & feature extraction approaches, the second is feature selection by using chi-square approach,and finally classification by using Logistic Regression Algorithm approaches. Some techniques are illustrated in details while the others with general explanation. In this research we used two datasets (dataset1 & dataset2) in order to test and train our program. The final result after applying chi-square and logistic regression algorihm to the aforementioned data set , the category of each input case will be determines by the likelihood of belonging to one of the two categories in the data set. [ABSTRACT FROM AUTHOR]

In this paper, texture classification is applied on coin image using coin recognition system. This system consists of training and testing phase. Based on Hough transform, firstly coin position will be identified in training phase. After the identification of the position, the texture feature will be extracted and uses these features to generate the dictionaries based on the Bag of word (BoW) approach. After extracting the texture feature, the image is divided into rings and fans structure and it follows the rotation invariant property. Further the features are quantized and generated a histogram to represent the texture features of the coin. In testing phase, the Support Vector Machine (SVM) is used to train a model for each class of the coin. Experimental results show that the proposed system provides more accuracy of recognition rate when compared to related studies. [ABSTRACT FROM AUTHOR]

The description of data structure for the description of geometrical three-dimensional object model is presented in the paper. Use of such data structure allows to develop algorithms of models creation for complex geometrical objects. The present possibility has essential value at creation of geometrical modeling system which can form a basis for development of specialized systems of automation, for example, of design and technological preparation of production. [ABSTRACT FROM AUTHOR]

This paper explores feature extraction algorithms for crack characterization in eddy current inspection of fastener sites. A novel feature extraction method fitting approximate models to data associated with geometric part features addressing adjacent fastener sites and panel edges is developed. Data classification methods in the circumferential direction around fastener sites are developed to better characterize fatigue cracks with improved noise invariance. Model-assisted probability of detection results are presented highlighting the benefit of automation in NDE. © 2007 American Institute of Physics [ABSTRACT FROM AUTHOR]

Fuzzy Topographic Topological Mapping (FTTM) is a model for solving neuromagnetic inverse problem. FTTM consists of four components and connected by three algorithms. FTTM version 1 and FTTM version 2 were designed to present 3D view of an unbounded single current and bounded multicurrent sources, respectively. In 2008, Suhana introduced some definitions on sequence ofFTTM. One ofthe features produced from the sequences ofFTTM is Cube of FTTM. A cube ofFTTM is a combination of two or more FTTM in FTTM«. In this paper, cube of finite vertices ofFTTM, namely FKn are discussed. Consequently, some theorems are proven in order to describe patterns for sequence of cubes for FKn based on this k-Fibonacci sequence. Interestingly, the cube oí FKn appears to be an example ofgeneralized Fibonacci sequence, namely the k-Fibonacci sequence. [ABSTRACT FROM AUTHOR]

Tomato is an important vegetable crop which is widely distributed due to its higher consumption rate. During the growing period, this plant is attacked by various diseases from bacteria, fungi and viruses. Diseases that attack can affect growth conditions in plants. This disease can be recognized visually because it has a unique color and texture characteristics. There are still deficiencies in visual recognition, it is the difficulties to recognize the similarities between one type of disease and another. In this study, the writers propose the K-Nearest Neighbor (K-NN) algorithm and feature extraction for the classification of tomato leaf disease images. Our research shows that the KNN algorithm is able to recognize and classify tomato leaf diseases effectively with an accuracy of 96.58%. [ABSTRACT FROM AUTHOR]

In trend, the breakthrough of various software's and technologies to make betterment of photographs has led to misuse of the discoveries. In turn, there is a higher chance of destroying the actual view of the image and produce duplications. In the proposed system, using Speeded Up Robust Feature (SURF) algorithm to abstract feature points on the images and match them to detect the occurrence of fraudulent in the image. SURF works in three phases. Firstly, the images are scanned, features are extracted, followed by feature description and finally the features are matched. Fraudulent is detected with the help of matched Features. Further to study the efficiency of algorithm, SURF and SIFT are implemented and the time taken to process the image is calculated. SURF seems to be faster than SIFT in processing and feature matching which speeds up forgery detection. Thus, SURF is speeder than SIFT. [ABSTRACT FROM AUTHOR]

This paper explores improving crack detection capability for Giant Magnetoresistive (GMR) measurements through simulation and development of automated data analysis algorithms. Simulated studies were performed that highlight differences in the measurements for varying fastener material properties and indicate potential locations for greater sensitivity to cracks. Automated algorithms were demonstrated to register fastener locations and extract crack features. To address the particular challenge of crack detection at fasteners near part edges, a model-based approach is introduced. [ABSTRACT FROM AUTHOR]

The paper describes a method for processing data from an ultrasonic transducer array. The proposed algorithm is formulated in such a way that it is reversible, i.e. the raw data set can be recovered from the image. This is of practical significance because it allows the raw-data to be spatially filtered using the image to extract, for example, only the raw data associated with a particular reflector. The method is tested on experimental data obtained from a commercial 64-element, 5-MHZ array on an aluminium specimen that contains a number of machined slots and side-drilled holes. The raw transmitter-receiver data corresponded to each reflector is extracted and the scattering matrices of different reflectors are reconstructed. It allows the signals from 1-mm-long slot and a 1-mm-diameter hole to be clearly distinguished and the orientation and the size of the slots to be determined. [ABSTRACT FROM AUTHOR]

We describe a specific problem of extraction a small subset E of r output features (described by matrixes Emxr ∈ Mmxn) in Euclidean space, on the basis of original data represented by input matrix S ∈ Mmxn(ℝ), where m ≪ n. Upon reducing S to Emxr, which is a continuation of [1], [2] and [3], we enforce r ≪ n and r < m. Reduced matrix Emxr encodes essential features in input dataset (see e.g. [6], [7] and [8]). In order to reduce the size of S we determine r hidden attributes by applying here a new algorithm with non-uniform density of attributes δLnum ( fi), for each class separately. Finally, this procedure generates a small subset C(Er) of representative and essential attributes, illustrated graphically, with different properties. [ABSTRACT FROM AUTHOR]

A quantitative assessment protocol was developed to evaluate potential signal processing methods for THz NDE of spray-on foam insulation. Automated feature extraction methods were developed to replicate the important visual classification characteristics used by inspectors and a new merit assessment approach was developed to identify the most promising combination of signal processing algorithms. Through application of the protocol to test data, several signal processing methods were found to improve the discrimination of defects. © 2007 American Institute of Physics [ABSTRACT FROM AUTHOR]